China Restricts OpenClaw AI at Banks, State Agencies Over Security Concerns

Chinese authorities have restricted state-owned enterprises and government agencies from running OpenClaw AI applications on office computers, moving to address potential security risks as the agentic AI tool gained rapid adoption across the country. Government agencies and state-owned enterprises, including China's largest banks, received notices in recent days warning against installing OpenClaw software on office devices. Several agencies were instructed to notify superiors if employees had already installed the apps, triggering security checks and possible removal. Some employees and families of military personnel were banned from installing the software on office or personal devices connected to company networks. Other notices required prior approval rather than an outright ban. OpenClaw is an open-source AI agent that automates daily tasks including email processing, schedule management, and travel booking. Beijing's concern centers on the platform's requirement for unusually broad access to private data and its ability to communicate externally, which could expose government computers to attack. The restrictions contrast with enthusiasm elsewhere in China's tech sector. Shenzhen's Longgang district separately announced subsidies to support OpenClaw industry development, reflecting conflicting regional approaches to the technology. Tencent and Zhipu have also launched AI agents built on OpenClaw.